i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Effectiveness factor correlations from simulations

of washcoat nickel catalyst layers for small-scale
steam methane reforming applications

Areum Jeong a, Dongwoo Shin a, Seung Man Baek b, Jin Hyun Nam c,*
a
School of Mechanical and Aerospace Engineering, Seoul National University, Seoul, 08826, Republic of Korea
b
R & D Center, Samsung Display Co. Ltd., Yongin, 17113, Republic of Korea
c
School of Mechanical Engineering, Daegu University, Gyungsan, 38453, Republic of Korea

article info abstract

Article history: In this paper, the effectiveness factors of Ni/MgAl2O4 washcoat catalyst layers under steam
Received 1 May 2018 methane reforming (SMR) conditions relevant to small-scale hydrogen production systems
Received in revised form (1e3 bar pressure, 600e800
C temperature, and 2e4 steam-to-carbon ratio) are numeri-
7 June 2018 cally investigated. The effects of the washcoat properties, including the layer thickness (20
Accepted 8 June 2018 e80 mm), the mean pore diameter (10e40 nm), and the volume-specific catalyst surface
Available online xxx area (1.1e3.3  107 m2/m3), are also considered. The simulation is conducted by fully
considering the intrinsic reaction kinetics (Xu and Froment model) and multicomponent
Keywords: mass diffusion (Maxwell-Stefan equation). The numerically obtained effectiveness factors
Steam methane reforming are presented as a function of the methane conversion ratio and effective Thiele moduli,
Small-scale reformer and simple correlation equations are proposed for easy evaluation of the effectiveness
Effectiveness factor factors.
Washcoat catalyst layer © 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Reaction/diffusion process
Correlation equation

the refining of petroleum oil. Recently, the fuel cell technology

Introduction
Steam reforming is the primary process in the industrial
(large-scale) production of hydrogen by the catalytic reaction
of light hydrocarbons (mainly methane) obtained from natu-
ral gas [1]. A mixture of hydrocarbons and steam is passed
through pipes containing nickel catalyst at high temperatures,
where the gas mixture is converted to carbon monoxide and
hydrogen. Then, the carbon monoxide is further converted to
carbon dioxide, producing additional hydrogen. Hydrogen has
been mainly used in the industrial synthesis of ammonia and
operated with hydrogen fuel has drawn considerable research
attention as an alternative power source that is clean and
renewable [2,3] Accordingly, the demand for small-scale (on-
site) production of hydrogen by reforming for fuel cell appli-
cations is increasing [4e8].
Nickel (Ni) is the most commonly used catalyst material for
the steam methane reforming (SMR) process because it is
cost-effective and efficient [9,10]. In Ni-based catalysts,
nanosized Ni particles are uniformly dispersed on porous
support structures made of alumina (Al2O3) or spinel

* Corresponding author. School of Mechanical Engineering, Daegu University, 15 Naeri-ri, Jinryang-eup, Gyungsan, 38453, Republic of
Korea.
E-mail address: jinhnam@gmail.com (J.H. Nam).
https://doi.org/10.1016/j.ijhydene.2018.06.059
0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
2 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

(MgAl2O4), in which the Ni content generally ranges from 7 to

Nomenclature 15 wt%. The SMR process with Ni catalysts is generally
described by global three-step reaction [11,12]; (1) steam
Acat Volume-specific catalyst surface area (m2 m3)
reforming, (2) water-gas shift, and (3) reverse methanation.
Ak Face area at grid point k (m2)
The overall process is strongly endothermic, and thus, suffi-
DCH4 Average methane diffusivity in the washcoat
cient heat should be provided, such as through the use of
layer (m2 s1)
multifunctional reactors [13] (composed of a reforming
Dij Binary diffusivity of species pair i  j (m2 s1)
reactor, an exothermic reactor, and a heat exchanger).
DiK Knudsen diffusivity of species i (m2 s1)
Two broad types of reformer designs are used for small-
dpore Mean pore diameter (m)
scale hydrogen production systems for fuel cell applications
Keq;m Equilibrium constant for reaction m (see
[14,15]. The first is packed-bed reformers [16e18], where
Appendix A for units)
porous Ni catalyst pellets are packed in a container and the
Ki Adsorption coefficient of species i (see
steam/methane gas mixture flows through it. In these re-
Appendix A for units)
formers, relatively large size of pellets results in low catalyst
Kv Viscous flow permeability (m2)
utilization. In addition, heat transfer is often limited by the
km Reaction rate constant for reaction m (see
low thermal conductivity of the catalyst pellets. The second
Appendix A for units)
type of the reformer design is coated-wall reformers [19e25],
Lcat Thickness of washcoat catalyst layer (m)
where a thin catalyst layer (called the washcoat) is formed
Mi Molecular mass of species i (kg kmol1)
directly on the reactor walls. Typical coated-wall reformers
Mwc;CH4 Nominal methane diffusion rate (mol m2 s1)
include catalytic plate reformers (also called microchannel
N Number of volume cells
reformers) [19e21] and monolithic reformers [22e25]. In cat-
Ni Molar flow rate of species i (mol s1)
alytic plate or microchannel reformers, the channels for gas
ni Molar flux of species i (mol m2 s1)
combustion are placed next to the channels for the SMR pro-
pi Partial pressure of species i (Pa)
cess, which provides an efficient heat supply for reforming. In
pt Total pressure of species i (Pa)
general, coated-wall reformers exhibit lower flow resistance
Rg Universal gas constant (8.314 J mol1 K1)
and higher catalyst utilization than packed-bed reformers.
Rwc;m Nominal reaction rate of reaction m (mol m2
The effectiveness factor is a quantitative measure of the
s1)
utilization of catalyst materials and is defined as the ratio of
rm Reaction rate for reaction m (kmol kg1 1
cat h )
actual overall rate of reaction to the rate of reaction that would
SC Steam-to-carbon ratio
result if entire interior surface were exposed to the external
si Volumetric source for species (mol m3 s1)
washcoat surface conditions [26,27]. The active surfaces of
T Temperature (K)
porous catalyst pellets and washcoat layers cannot be fully
Vk Cell k, or its cell volume (m3)
utilized due to the relatively slow mass transport through the
vi Diffusion volumes of species i
pores. The effectiveness factors of Ni catalyst pellets under
xi Mole fraction of species i
industrial (large-scale) SMR conditions have long been studied
z Coordinate (m)
by many researchers [12,28e30] because these factors are
zk Grid point k, or its coordinate location (m)
essential information for the proper design of SMR reactors.
Dzk Distance between two adjacent cell centers at
The authors have recently studied the effectiveness factors of
grid point k (m)
Ni catalyst pellets under small-scale SMR conditions near at-
Greek symbols mospheric pressure and then proposed simple correlation
bCH4 Methane conversion ratio equations [31,32]. However, few references that can be used
eq
bCH4 Equilibrium methane conversion ratio for estimating the effectiveness factors of Ni washcoat cata-
ε Porosity lyst layers under small-scale SMR conditions are currently
fm Effective Thiele modulus for reaction m available [19,33].
f*m Modified Thiele modulus for reaction m Thus, this study numerically investigated the effectiveness
hm Effectiveness factor for reaction m factors of thin washcoat catalyst layers made of Ni/MgAl2O4 in
mmx Viscosity of gas mixture (kg m1 s1) the SMR process. The reforming conditions for small-scale
rcat Apparent catalyst density (kg m3) hydrogen production systems, such as a pressure range of
t Tortuosity 1e3 bar, a temperature range of 600e800
C, and a steam-to-
carbon ratio (S/C) range of 2e4, were considered. In addition,
Superscript/Subscripts
the effects of the washcoat properties, including the layer
∞ Flow stream condition
thickness (20e80 mm), the mean pore diameter (10e40 nm),
eff Effective value
and the volume-specific catalyst surface area
I, II, III Reaction index
(1.1e3.3  107 m2/m3), were also considered. For this purpose,
i; j Species index
a one-dimensional, isothermal reaction/diffusion process was
k Grid or cell index
simulated inside the washcoat layer by considering the
o Inlet or initial value
intrinsic reaction kinetics for the steam reforming, water-gas
shift, and reverse methanation reactions (Xu and Froment
model [11,12]) and multicomponent mass diffusion (Maxwell-
Stefan equation [34,35]). The standard composition of

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 3

reformate gas was first determined using the same reaction explains the SMR process occurring inside a channel. As the
kinetics and then applied to the surface of the washcoat layers steam/methane mixture gas flows through the channel,
as boundary conditions. The obtained effectiveness factors methane (CH4) and water vapor (H2O) diffuse into the wash-
were presented as a function of the methane conversion ratio coat layer. These reactant gases are converted to hydrogen
and effective Thiele moduli. Predictive correlation equations (H2), carbon monoxide (CO), and carbon dioxide (CO2) by the
were also proposed for easy utilization of the effectiveness SMR process inside the washcoat layer. Finally, the product
factors determined here in designing reformers or performing gas species diffuse back into the reforming gas stream.
complex numerical simulations. The gas composition in the washcoat layer is different
from that in the flow channel due to the occurrence of the SMR
reactions. The reactant concentrations are lower in the
Theory and calculation washcoat layer than in the channel, whereas the product
concentrations are higher. This difference causes the reaction
Physical model rate of the SMR process to be lower than the reaction rate
calculated with the bulk species concentrations in the chan-
Fig. 1 illustrates a schematic of a monolithic catalyst nel, resulting in lower effectiveness factors. The concentra-
composed of a solid substrate structure with many embedded tion difference also explains why lower catalyst effectiveness
channels. The channel walls are coated with a thin layer of factors are obtained when the product concentrations are
active catalyst materials called the washcoat layer. Fig. 1b high (for example, near the inlet of the reformers) and/or
when the washcoat layer is thick and dense.
Fig. 1c shows the one-dimensional (1D) domain of the
washcoat layer considered in this study, where z axis is
st
ataly oriented from the layer adherent to the monolith surface
C
(a) nolith towards the space for gas flow stream. Inside the 1D
Mo domain, N volume cells (V1 ; V2 ; /; VN ) are defined by N þ 1
grid points (z0 ; z1 ; /; zN ). In addition, two volume-less cells
(V0 ; VNþ1 ¼ 0) are also defined at the two boundaries at z ¼ 0
Channel and z ¼ Lcat (Lcat is the washcoat layer thickness). The grid
points were nonuniformly generated according to
Substrate zk ¼ ðk=NÞ0:5 Lcat so that more volume cells were located near
the channel (z ¼ Lcat ). The partial pressure, pi , and reaction
Washcoat
rates, rI , rII , and rIII , were calculated at the volume cells,
while the molar flux, ni , was calculated at N grid points (the
interface between cells at z1 ; z2 ; /; zN ). A total of 50 volume
cells (N ¼ 50) was sufficient to obtain converged results. The
details of the numerical model are explained below. Heat
(b) ts transfer was not considered in this study by assuming the
uctan
d CO 2 uniform temperature inside the thin washcoat layer. In fact,
Pro CO Baek et al. [31] showed that the temperature distribution
s H2
ant was negligible inside a spherical catalyst pellet (3 mm in
eact H O
R eam diameter).
1D Domain

Str
2
CH 4
s
z Ga
ng Governing equations
r mi
Refo
In simulating the SMR process, we should consider five gas
Washcoat species, methane (species 1: CH4), water vapor (2: H2O),
Substrate hydrogen (3: H2), carbon monoxide (4: CO), and carbon dioxide
(5: CO2). For the 1D calculation, the conservation of gas species
(c) Reaction Rate
i in the washcoat layer is expressed as
Channel

+2 Volumeless Cells
z0 = 0 r I , r II , r III N Volume Cells dni
¼ si ðfor i ¼ 1; 2; /; 5Þ; (1)
dz
ni, k 1 ni, k
where ni is the molar flux (mol m2 s1) and si is the volumetric
pi, k 1 pi, k pi, k +1 z generation (mol m3 s1) of species i. The calculation of ni and
Substrate

zk si is explained in the following sections.

i = 1 : CH 4 4 : CO zk 1 zk +1 zN = Lcat The boundary conditions employed in this study are
2 : H 2O 5 : CO 2 k - th Finite Volume
expressed as
3 : H2 Vk = z k − z k 1
ni ¼ 0 at z ¼ 0; (2)
Fig. 1 e Schematic of (a) a monolithic-type coated wall
reformer, (b) the SMR process occurring inside a channel,
pi ¼ pi;∞ ¼ xi;∞ pt;∞ at z ¼ Lcat : (3)
and (c) the calculation grid used in the simulation.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
4 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

Reaction kinetics where Kv is the viscous flow permeability (m2); mmx is the vis-
cosity of the gas mixture (kg m1 s1), which is the molar
This study considers the global three-step SMR reaction average of the species viscosities [36]; Deffij is the effective bi-
expressed as nary diffusivity (m2 s1) for species pair ij; Deff
iK is the effective
I. steam reforming: Knudsen diffusivity (m2 s1) of species i; and pt is the total
pressure (Pa).
CH4 þ H2 O4CO þ 3H2 ; Dh~298:15K ¼ þ206 kJ mol ;
1
(4)
The effective binary diffusivity, Deff
ij , is determined as
II. water-gas shift: ε
ij ¼ Dij ;
Deff (17)
t
CO þ H2 O4CO2 þ H2 ; Dh~298:15K ¼ 41 kJ mol ;
1
(5)
where ε and t are the porosity and tortuosity of the washcoat
III. reverse methanation:
layer. Dij is the binary diffusivity for species pair ij, calculated
CH4 þ 2H2 O4CO2 þ 4H2 ; Dh~298:15K ¼ þ165 kJ mol :
1 as (pt with units of bar)
(6)

The intrinsic kinetics for the above three reactions were 0:00143T1:75
Dij ¼ 104 h i2 ; (18)
determined using the rate equations proposed by Xu and 1=2 1=3 1=3
pt Mij vi þ vj
Froment [11,12]. The reaction rates for reactions I, II, and III, rI ,
rII , and rIII (kmol kg1 1
cat h ), are expressed as where vi and vj denote the diffusion volumes defined by Fuller
!, et al. [37] (25.14 for CH4, 13.1 for H2O, 6.12 for H2, 18.0 for CO,
kI p3H2 pCO and 26.7 for CO2) and Mij is the mixture molecular mass
rI ¼ pCH4 pH2 O  DEN2 ; (7)
p2:5
H2 Keq;I defined as

      1
kII pH pCO2 Mij ¼ 2 ð1=Mi Þ þ 1 Mj : (19)
rII ¼ pCO pH2 O  2 DEN2 ; (8)
pH2 Keq;II
The molecular masses of CH4, H2O, H2, CO, and CO2 are
!, 16.043, 18.015, 2.016, 28.01, and 44.01 kg kmol1, respectively
kIII p4H2 pCO2 [37].
rIII ¼ pCH4 p2H2 O  DEN2 ; (9)
p3:5
H2 Keq;III The effective Knudsen diffusivity is determined as

where DEN is defined as sffiffiffiffiffiffi

ε ε 97 T
iK ¼ DiK ¼
Deff dpore ; (20)
KH2 O pH2 O t t 2 Mi
DEN ¼ 1 þ KCO pCO þ KH2 pH2 þ KCH4 pCH4 þ : (10)
pH2
where dpore is the mean pore diameter and Mi is the molecular
The partial pressures in Eqs. (7)e(10) have units of bar. In mass (kg kmol1) of species i.
Appendix A, the pre-exponential factors, activation energies,
heats of adsorption, and equilibrium constants are summarized. Effectiveness factor
The volumetric source for species i, si (mol m3 s1), in Eq.
(1) is calculated from the reaction rates given in Eqs. (7)(9) The catalyst effectiveness factors for the steam reforming
and the stoichiometry of the SMR reactions in Eqs. (4)(6) (reaction I), hI , and the reverse methanation (reaction III), hIII ,
and expressed as are evaluated as

s1 ¼ rcat ð  rI  rIII Þ=3:6; Z Z Lcat

(11) Lcat      
rcat rI pi i¼1/5 ; T∞ dz rcat rIII pi i¼1/5 ; T∞ dz
hI ¼ 0
  ; hIII ¼ 0
  ;
s2 ¼ rcat ð  rI  rII  2rIII Þ=3:6; (12)  
rcat Lcat rI pi;∞  ; T∞ rcat Lcat rIII pi;∞  ; T∞
i¼1/5 i¼1/5
(21)
s3 ¼ rcat ð þ 3rI þ rII þ 4rIII Þ=3:6; (13)
where rcat and Lcat are the apparent density and thickness of

s4 ¼ rcat ð þ rI  rII Þ=3:6; (14) the washcoat layer, respectively. The terms rI ðpi i¼1/5 ; T∞ Þ and

rIII ðpi i¼1/5 ; T∞ Þ denote the actual reaction rate inside the
s5 ¼ rcat ð þ rII þ rIII Þ=3:6; (15) washcoat layer due to the mixture partial pressures, pi . In
 
 
where the term 1/3.6 (¼1000/3600) is a factor to convert the contrast, rI ðpi;∞  ; T∞ Þ and rIII ðpi;∞  ; T∞ Þ denote the
i¼1/5 i¼1/5
unit of si from kmol m3 h1 to mol m3 s1.
nominal reaction rate under the conditions of the bulk flow
stream. In this study, we focused only on the effectiveness
Mass transport
factors of the steam reforming and reverse methanation re-
actions. Note that pi;∞ and T∞ are different from the average
The mass flux of each species in the washcoat layer is deter-
values in the bulk flow stream; instead, they should be un-
mined by the dusty-gas model [34,35], expressed as
derstood as the partial pressures and temperature in the flow
stream near the washcoat layer surface.
1 dpi pi Kv dpt X 5
xj ni  xi nj ni
  ¼ þ eff ðfor i ¼ 1;2;/;5Þ; In this study, the effective Thiele moduli for the steam
Rg T dz Rg T mmx DiK dz j¼1
eff
Deff DiK
ij reforming (reaction I), fI , and that for the reverse methanation
(16) (reaction III), fIII , are defined as [31,32].

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 5

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi The species mole fraction in the reformer, xi , also varies
Rwc;I Rwc;III along the axial direction according to the progress of the SMR
fI ¼ ; fIII ¼ ; (22)
Mwc;CH4 Mwc;CH4 process. The standard reformate gas composition was
numerically obtained as a function of bCH4 using the same
where Rwc;I and Rwc;III denote the nominal reaction rates of
reaction kinetics shown in Eqs. (7)(10) for all 27 simulated
reactions I and III and Mwc;CH4 denotes the methane diffusion
reforming conditions (pressures of 1, 2, and 3 bar; tempera-
rate into the washcoat layer. The nominal reaction rates are
tures of 600, 700, and 800
C; and S/C ratios of 2, 3, and 4).
defined as
Previously, the authors used polynomial correlations to eval-
 

 
Rwc;I ¼ rcat Lcat rI pi;∞  ; T∞ Rwc;III ¼ rcat Lcat rIII pi;∞  ; T∞ ;
i¼1/5 ; i¼1/5
(23)
and the methane diffusion rate is defined as
uate xi as a function of bCH4 [31,32]. However, in this study,
tabular mole fraction data (xi with respect to bCH4 ) were
directly used as the boundary conditions of Eq. (3).
Numerical calculation
DCH4 pt;∞  eq


Mwc;CH4 ¼ xCH4 :∞  xCH4 ; (24)
Lcat Rg T∞ For numerical calculation, the governing equations and
eq
where xCH4 is the equilibrium methane mole fraction of the boundary conditions were discretized based on the grid
SMR process and is dependent on reforming pressure, tem- structure shown in Fig. 1c. The governing equation of Eq (1).
perature, and S/C ratio (see Appendix B). for gas species i was discretized for N volume cells (V1 ; V2 ; /;
In Eq. (24), DCH4 is the average methane diffusivity in the VN ) as
washcoat layer evaluated with the bulk gas concentrations,
ni;k Ak  ni;k1 Ak1 ¼ si;k Vk ðfor k ¼ 1; 2; /; NÞ; (29)
defined as
where Ak and Ak1 denote the east and west face area of the Vk
1 1 1 volume cell, respectively (Ak ¼ 1 for 1D problems). The reac-
¼ þ ; (25)
DCH4 Deff
m;CH4 Deff
CH4 K tion rates, rI;k , rII;k , and rIII;k , in Eqs. (7)(9) and the resulting
where Deff species sources, si;k , in Eqs. (11)(15) were also calculated for
m;CH4 is defined as
these N volume cells. The boundary conditions of Eqs. (2) and
1  xCH4 ;∞ xH O;∞ xH2 ;∞ xCO;∞ xCO2 ;∞ (3) were implemented by the following equations for two
¼ eff 2 þ þ eff þ : (26)
Deff
m;CH4 DCH4 H2 O Deff
CH4 H2 DCH4 CO Deff
CH4 CO2 volume-less cells at the boundaries (V0 ; VNþ1 ), which are

At near-atmospheric pressures, Deff CH4 K is much smaller pi;k ¼ pi;kþ1 ðfor k ¼ 0Þ; (30)
than Deff
m;CH4 due to the small size of pores in the washcoat
layer. In this case, DCH4 can be approximated as pi;k ¼ xi;∞ pt;∞ ðfor k ¼ N þ 1Þ: (31)
DCH4 z0:97Deff
CH4 K at 1 bar, z0:93Deff CH4 K at 2 bar, and
z0:90DCH4 K at 3 bar [32].
eff Thus, the species partial pressure at each cell center is
the primary solution variable (pi;k for i ¼ 1; 2; /; 5 and k ¼ 0;
Standard gas composition 1; /; N; N þ 1).
From Eq. (16), the molar flux of species i, ni;k , was calculated
Consider a situation in which steam/methane mixture gas at the grid point zk by solving the following discretized equa-
flows through a long microchannel reformer whose wall is tion, written as
coated with a washcoat catalyst layer and is operated at a ~i;k
1 pi;kþ1  pi;k p Kv pt;kþ1  pt;k
constant pressure and temperature. The S/C ratio, SC, is  
Rg T Dzk Rg T mmx Deff
iK
Dzk
calculated from the molar flow rates of methane, NoCH4 (mol
X
5
x~j;k ni;k  x~i;k nj;k ni;k
s1), and water vapor, NoH2 O , at the inlet, expressed as ¼ þ ðfor k ¼ 1; 2; /; NÞ; (32)
j¼1
Deff
ij Deff
iK
NoH2 O
SC ¼ : (27) where Dzk denotes the distance between the centers of two
NoCH4
adjacent volume cells (Vk and Vkþ1 ) and calculated as Dzk ¼
Then, SC is uniform inside the reformer provided that no ~i;k is the interpolated species
ðzkþ1  zk1 Þ=2. In addition, p
gas is injected into or extracted from the flow stream. partial pressure at the grid point zk , and x~i;k and x~j;k are the
The amount of methane in the reformer decreases along interpolated species mole fractions. The molar flux, ni;0 , at z ¼
the axial direction according to the SMR process. The methane 0 was set to zero according to Eq. (2).
conversion ratio, bCH4 , is a measure of the progress of the SMR The nonlinear equations of Eqs. (29)(32) were solved using
process, defined as a nonlinear equation solver called the engineering equation
NCH4 solver (EES) [38]. In the calculation, a grid number of 50 (N ¼
bCH4 ¼ 1  ; (28) 50) was sufficient to obtain accurate results. The simulated
NoCH4
SMR conditions and washcoat properties are summarized in
where NCH4 denotes the molar flow rate of methane at a Table 1, where the underlined values denote the base-case
certain downstream position. In the reformer, bCH4 is 0 at the conditions. Note that the porosity, ε, and tortuosity, t, are
inlet, increases along the axial direction, and eventually rea- set to 0.5 and 4.0, respectively, by referencing to Xu and Fro-
eq
ches the equilibrium value, bCH4 , corresponding to the given ment (ε ¼ 0:528, t ¼ 3:54) [11,12] and Zanfir and Gavriilidis (ε ¼
reforming pressure, temperature, and S/C ratio. 0:4, t ¼ 4) [19].

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
6 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

0.6
Results and discussion Base Case Condition & (a)
CH4 = 0.3 Washcoat Properties
0.5
1 bar, 700°C , SC = 3

Mole Fraction, xi
Base-case results xH 2
0.4
Fig. 2 shows the distribution of the species mole fraction, xi ,
0.3 xH 2 O
inside the washcoat catalyst layer during the SMR process.
The results were obtained with the base-case reforming con- 0.2
ditions and washcoat properties provided in Table 1. The
xCO 2
methane conversation ratio, bCH4 , varies from 0.3 in Fig. 2a to 0.1 xCO
0.6 in Fig. 2b and 0.9 in Fig. 2c, and accordingly, the species xCH 4
0.0
concentrations at the boundary (z ¼ 50 mm) are different. In
0.6
Fig. 2, the mole fraction of methane, xCH4 , gradually decreases
= 0.6 (b)
from z ¼ 50 mm towards z ¼ 0 mm, indicating that the 0.5
CH 4
xH 2
reforming reactions occur throughout the thin washcoat

Mole Fraction, xi
layer. However, the reaction rate is higher near z ¼ 50 mm, as 0.4
observed by the higher gradient of xCH4 at the boundary. For
eq
bCH4 ¼ 0:9 (close to bCH4 ¼ 0:9717 at 1 bar, 700
C, and S/C ratio 0.3 xH 2 O
of 3), the variation in the mole fraction is smaller for all five
0.2
species, which implies that the reforming reactions become
xCO
slower near the equilibrium conditions. 0.1
Fig. 3 presents the effectiveness factors, hI for steam xCH 4 xCO 2
reforming, hII for water-gas shift, and hIII for reverse metha- 0.0
nation, as a function of bCH4 , obtained with a pressure of 1 bar 0.6
(c)
and the base-case washcoat properties. A total of 9 reforming xH 2
conditions were simulated by varying the reforming temper- 0.5 = 0.9
CH 4
Mole Fraction, xi

ature (T∞ ¼ 600; 700, and 800
C) and S/C ratio (SC ¼
0.4
2; 3; and 4). As shown in Fig. 3a and c, hI and hIII are very small
near bCH4 ¼ 0 and increase as bCH4 increases. This trend can be 0.3 xH 2 O
understood by inspecting Fig. 2. At low bCH4 , the reaction rates,
rI and rIII , are generally high due to high reactant concentra- 0.2
tions, which leads to considerable difference in the species xCO
concentrations between in the washcoat layer and in the bulk 0.1
xCH 4 xCO 2
flow stream, as shown in Fig. 2a. On the contrary, at high bCH4 0.0
eq
near bCH4 , the reaction rates become lower and thus the con- 0 10 20 30 40 50
centration difference diminishes, as shown in Fig. 3c. Recall Washcoat Coordinate, z ( m)
that, in Eq. (21), the effectiveness factor is defined as the ratio
Fig. 2 e Distribution of the species mole fraction, xi , in the
washcoat catalyst layer at the base-case conditions: (a)
bCH4 ¼ 0:3, (b) bCH4 ¼ 0:6, and (c) bCH4 ¼ 0:9.
Table 1 e The reforming conditions and washcoat layer
properties considered in this study.
Parameters Values of the actual total reaction rate (obtained by integrating vari-
Reforming conditions
able reaction rates inside the washcoat layer) to the nominal
Temperature, T∞ 600, 700a, 800
C total reaction rate (obtained by multiplying the total catalyst
Pressure, pt;∞ 1, 2, 3 bar surface area with a single reaction rate due to the bulk flow
Steam-to-carbon ratio, SC 2, 3, 4 conditions). Thus, the effectiveness factors, hI and hIII , are
Washcoat layer properties initially low at low bCH4 and increase with bCH4 .
Catalyst density, rcat 2,355b kg/m3
In general, conditions more favorable for enhancing the
Volume-specific catalyst surface area, 1.1, 2.2b, 3.3  107 m2/m3
reaction rate than the diffusion rate result in larger concen-
Acat
Layer thickness, Lcat 20, 50, 80 mm tration difference between in the washcoat layer and in the
Porosity, ε 0.5 bulk flow stream, leading to lower effectiveness factors.
Tortuosity, t 4.0 Likewise, conditions more favorable for enhancing the diffu-
Mean pore diameter, dpore 10, 25, 40 nm sion rate lead to higher effectiveness factors. In Fig. 3a and c, hI
Viscous permeability, Kv 1  1018 m2 and hIII decrease as the reforming temperature increases.
a
The underlined values are the base-case conditions and wash- Higher temperature increases both the reaction rate and the
coat properties. diffusion rate; however, the degree of reaction rate enhance-
b
The volume-specific catalyst surface area, Acat , of 2.2  107 m2/ ment is higher than that of diffusion rate enhancement, and
m3corresponds to the apparent catalyst density, rcat , of 2355 kg/
thus the effectiveness factors decrease with the reforming
m3, from Xu and Froment [11,12] (2355 kg/m3  9300 m2/
temperature. The dependency of hI and hIII on the S/C ratio
kgcat ¼ 2.2  107 m2/m3).

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 7

0.6
Base Case Properties (a)
0.6 Fixed @ pt , = 1 bar
0.5
Base Case Properties (a) T = 600 C

I
0.5 Fixed @ pt , = 1 bar

Effectiveness Factor,
0.4
I

Steam/Carbon Ratio
Steam/Carbon Ratio
Effectiveness Factor,

0.4 SC = 2:
3: 0.3
SC = 2:
T = 600 C 3:
4: 700 C 4:
0.3
0.2
700 C
800 C
0.2
0.1

0.1 800 C 0
0 4 8 12 16 20
0 Effective Thiele Modulus, I
0 0.2 0.4 0.6 0.8 1 0.6
Methane Conversion, CH 4
Base Case Properties (b)
100 Fixed @ pt , = 1 bar
0.5
Base Case Properties (b) T = 600 C

III
Fixed @ pt , = 1 bar
Effectiveness Factor, 0.4
II

Steam/Carbon Ratio Steam/Carbon Ratio

Effectiveness Factor,

10 SC = 2: SC = 2:
0.3
3: 3:
800 C 4: 700 C 4:
0.2

1
0.1 800 C
T = 600 C

0
700 C 0 4 8 12 16 20

0.1 Effective Thiele Modulus, III

0 0.2 0.4 0.6 0.8 1
Methane Conversion, Fig. 4 e Variations in the effectiveness factors with respect
CH 4
to the effective Thiele moduli, fI and fIII , at the base-case
0.6
conditions: (a) hI for steam reforming and (b) hIII for reverse
Base Case Properties (c)
methanation.
0.5 Fixed @ pt , = 1 bar
III
Effectiveness Factor,

0.4 Steam/Carbon Ratio

SC = 2:
3: seems to be relatively small when the reforming temperature
0.3
T = 600 C 4: is higher than 700
C.
In Fig. 3b, hII for the water-gas shift reaction shows a trend
0.2 quite different from those of hI and hIII ; hII is initially very high
700 C
for a small bCH4 and then decreases as bCH4 increases. An
effectiveness factor hII above one may be explained as follows.
0.1
For a small bCH4 , the amount of carbon monoxide in the bulk
800 C
gas stream is small, leading to a low rate of water-gas shift
0 reaction. However, the amount of carbon monoxide in the
0 0.2 0.4 0.6 0.8 1
washcoat layer is larger than that in the bulk gas stream due to
Methane Conversion, CH 4 the reforming reactions. This results in a higher rate for the
water-gas shift reaction in the washcoat layer than that in the
Fig. 3 e Variations in the effectiveness factors with respect
bulk gas stream, leading to an hII value above one. We do not
to the methane conversion ratio, bCH4 , at the base-case
consider the effectiveness, hII , for the water-gas shift reaction
conditions: (a) hI for steam reforming, (b) hII for water-gas
in the following sections. However, considering the small ef-
shift, and (c) hIII for reverse methanation.
fects of the water-gas shift reaction on the SMR process, it is
the authors' opinion to set hII to around 0.5e1.0 for simulating

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
8 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

small-scale steam methane reformers with washcoat catalyst methanation are competing reactions that consume methane,
layers. a condition favorable for the former is usually unfavorable for
the latter.
Effects of reforming conditions As can be inferred from Fig. 4a and b, simple correlations
cannot predict hI and hIII for the various reforming tempera-
In Fig. 4, the effectiveness factors hI and hIII are plotted as a tures and S/C ratios considered in this study. Thus, modified
function of the corresponding effective Thiele moduli, fI and Thiele moduli, fI and fIII , are introduced to make the effec-
fIII , defined in Eq. (22). A total of nine simulation cases are tiveness factor data fall on a single correlation curve [32].
plotted by varying the reforming temperature (T∞ ¼ 600; 700, These modified Thiele moduli are defined as
and 800
C) and S/C ratio (SC ¼ 2; 3; and 4). In Fig. 4, the  1:3  0:07  0:9  0:6
effectiveness factors monotonically decrease as the effective T∞ SC T∞ SC
fI ¼ fI ; fIII ¼ fIII : (33)
Thiele modulus increases. In general, higher effectiveness 700
C 3 700
C 3
factors are obtained for lower reforming temperatures The effectiveness factors shown in Fig. 4 are plotted again
(accordingly, when the reaction rate is lower). In Fig. 4a, a in Fig. 5, this time with respect to the modified Thiele moduli,
higher S/C ratio generally leads to a lower effectiveness factor fI and fIII .
for the steam reforming reaction, hI . In contrast, a higher S/C Fig. 5 clearly shows that the effectiveness factors for
ratio results in a higher effectiveness factor for the reverse various reforming temperatures and S/C ratios can be fitted by
methanation reaction, hIII , as shown in Fig. 4b. This result may simple correlation equations, expressed as
be explained as follows: since steam reforming and reverse

0.6 0.8
Base Case Properties (a) Base Case Properties (a)
Fixed @ SC = 3 T = 600 C
Fixed @ pt , = 1 bar Pressure
0.5
pt, = 1 bar:
I
I

T = 600 C 0.6 2 bar:

Effectiveness Factor,
Effectiveness Factor,

0.4
Steam/Carbon Ratio 3 bar:
SC = 2:
3: 700 C
0.3 4: 0.4
700 C
1.5
0.2 I = , I 1
1 + 0.41( I* )1.88
0.2
800 C Correlation
0.1
800 C

0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
I
Methane Conversion, CH 4
0.6 0.8
Base Case Properties
(b)
Base Case Properties (b)
Fixed @ pt , = 1 bar Fixed @ SC = 3 T = 600 C Pressure
0.5
T = 600 C pt, = 1 bar:
III
III

0.6 2 bar:
Effectiveness Factor,

Steam/Carbon Ratio
Effectiveness Factor,

0.4 3 bar:
SC = 2:
3:
4:
0.3 0.4 700 C
700 C
1.5
0.2 III = , η III 1
* )1.03
1 + 4.14( III
0.2
Correlation
0.1 800 C
800 C

0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
III
Methane Conversion, CH 4

Fig. 5 e Variations in the effectiveness factors with respect Fig. 6 e Variations in the effectiveness factors with respect
to the modified Thiele moduli, fI and fIII , at the base-case to the methane conversion ratio, bCH4 , for reforming
conditions: (a) hI for steam reforming and (b) hIII for reverse pressures of pt;∞ ¼ 1, 2, and 3 bar: (a) hI for steam reforming
methanation. and (b) hIII for reverse methanation.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 9

decreases. The steam reforming and reverse methanation

1:5 1:5 reactions are retarded at higher reforming pressures, and
hI ¼  1:88 ðhI  1Þ; hIII ¼  1:03 ðhIII  1Þ:
1 þ 0:41 fI 1 þ 4:14 fIII thus, higher hI and hIII values are obtained.
To consider the reforming pressure effect, the modified
(34)
Thiele moduli are defined as
It should be noted that the modified Thiele moduli are not
1:3  0:07 
 0:18
derived on a theoretical basis. Instead, they are simply intro- T∞ SC pt;∞
fI ¼ fI

;
duced to efficiently correlate the effectiveness factors for 700 C 3 1bar
 0:9  0:6  0:2 (35)
various reforming conditions. However, Eqs. (33) and (34) can T∞ SC pt;∞
fIII ¼ fIII :
cover almost all the situations encountered during the SMR 700
C 3 1bar
process under conditions of 1 bar, 600e800
C, and a S/C ratio
Fig. 7 shows the effectiveness factors as a function of the
of 2e4.
modified Thiele moduli, fI and fIII , defined by Eq. (35). Clearly,
The effect of the reforming pressure, pt;∞ , is investigated in
by using these modified Thiele moduli, the effectiveness fac-
Fig. 6, where the effectiveness factors hI and hIII are presented
tors at various reforming pressures and temperatures are
as a function of the methane conversation ratio, bCH4 . A total
properly described by the correlation equations provided in
of 9 simulation cases are plotted for reforming pressures of 1,
Eq. (34). Thus, the present correlation approach is believed to
2, and 3 bar and temperatures of 600, 700, and 800
C at a fixed
be useful for retrieving the effectiveness factor data for
S/C ratio of 3. In Fig. 6, higher reforming pressures increase hI
eq various reforming conditions.
and hIII while the equilibrium methane conversion ratio, bCH4 ,

0.8 1
Base Case Properties (a) Fixed @ pt , = 1 bar, SC = 3 (a)
Fixed @ SC = 3
0.8 Washcoat Thickness
I

Lcat = 20 m:
I

0.6 T = 600 C
T = 600 C
Effectiveness Factor,

50 m:
Effectiveness Factor,

Pressure
pt, 80 m:
= 1 bar: 0.6
2 bar: 700 C
3 bar:
0.4 700 C
0.4 800 C
1.5
I = , I 1
800 C 1 + 0.41( I* )1.88
0.2
Correlation 0.2

0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
I
Methane Conversion, CH 4
0.8 1
Base Case Properties (b)
Fixed @ pt , = 1 bar, SC = 3 (b)
Fixed @ SC = 3 Washcoat Thickness
0.8 Lcat = 20 m:
III
III

0.6
T = 600 C
T = 600 C 50 m:
Effectiveness Factor,
Effectiveness Factor,

Pressure 80 m:
pt, = 1 bar: 0.6 700 C
2 bar:
0.4 3 bar:
700 C
0.4 800 C
1.5
III = , III 1
* )1.03
1 + 4.14( III
0.2
Correlation 0.2
800 C

0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
III
Methane Conversion, CH 4

Fig. 7 e Variations in the effectiveness factors with respect Fig. 8 e Variations in the effectiveness factors with respect
to the modified Thiele moduli, fI and fIII , for reforming to the methane conversion ratio, bCH4 , for washcoat layer
pressures of pt;∞ ¼ 1, 2, and 3 bar: (a) hI for steam reforming thicknesses of Lcat ¼ 20, 50, and 80 mm: (a) hI for steam
and (b) hIII for reverse methanation. reforming and (b) hIII for reverse methanation.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
10 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

Effects of washcoat properties moduli, fI and fIII . To account for the effect of the washcoat
thickness, the modified Thiele moduli are defined as
The structural properties of the washcoat catalyst layer also 1:3  0:07 
 0:37
significantly affect the effectiveness factors. Fig. 8 shows the T∞ SC Lcat
fI ¼ fI ;
effectiveness factors hI and hIII as a function of the methane 700
C 3 50mm
 0:9  0:6  0:15 (36)
conversion ratio, bCH4 . A total of 9 cases were simulated by T∞ SC Lcat
fIII ¼ fIII :
varying the reforming temperature, T∞ , between 600, 700, and 700
C 3 50mm
800
C and the washcoat layer thickness, Lcat , between 20, 50, Fig. 9 clearly shows that the correlation equations of Eq.
and 80 mm (the reforming pressure and S/C ratio are fixed at (34) adequately describe the effectiveness factors for various
1 bar and 3, respectively). As shown in Fig. 8, thinner washcoat washcoat thicknesses when the modified Thiele moduli
layers generally result in higher effectiveness factors, whereas defined in Eq. (36) are employed.
thicker layers result in lower factors. This difference is Next, the effect of the mean pore diameter, dpore , on the
because the reaction rate inside the washcoat layer increases washcoat effectiveness factors is investigated in Figs. 10 and
with the washcoat thickness while the diffusion rate de- 11. According to Eq. (20), the Knudsen diffusivity is propor-
creases. According to Fig. 8, the effectiveness factors can be as tional to the mean pore diameter. Due to the small pore
high as 0.8 for Lcat ¼ 20 mm when the reactions are very slow diameter of the washcoat layer, Knudsen diffusion becomes
(near equilibrium at T∞ ¼ 600
C). the dominant transport mechanism at near-atmospheric
The effectiveness factors, hI and hIII , shown in Fig. 8 are reforming pressures. For example, the effective Knudsen
replotted in Fig. 9, this time with respect to the modified Thiele diffusivity of CH4 with dpore ¼ 25 nm is 1.18  106 m2/s at T∞ ¼

1 0.8

Fixed @ pt , = 1 bar, SC = 3 (a) Fixed @ pt , = 1 bar, SC = 3 (a)

0.8 Pore Diameter

I

T = 600 C
I

T = 600 C Washcoat Thickness 0.6 dpore= 10 nm:

Effectiveness Factor,

25 nm:
Effectiveness Factor,

Lcat = 20 m:
50 m: 40 nm:
0.6 80 m:
700 C 0.4 700 C
0.4
1.5
800 C I = , I 1
1 + 0.41( I* )1.88
0.2
0.2 Correlation

800 C
0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
I
Methane Conversion, CH 4
1 0.8

Fixed @ pt , = 1 bar, SC = 3 (b) Fixed @ pt , = 1 bar, SC = 3 (b)

0.8 Pore Diameter

III
III

T = 600 C Washcoat Thickness 0.6 dpore= 10 nm: T = 600 C

Effectiveness Factor,

Lcat = 20 m: 25 nm:
Effectiveness Factor,

50 m: 40 nm:
0.6 80 m:
700 C
0.4
700 C
0.4
1.5
III = , III 1
* )1.03
1 + 4.14( III
0.2
0.2 Correlation
800 C

800 C
0 0
0 4 8 12 16 20 0 0.2 0.4 0.6 0.8 1
Modified Thiele Modulus, *
III
Methane Conversion, CH 4

Fig. 9 e Variations in the effectiveness factors with respect Fig. 10 e Variations in the effectiveness factors with
to the modified Thiele moduli, fI and fIII , for washcoat respect to the methane conversion ratio, bCH4 , for mean
layer thicknesses of Lcat ¼ 20, 50, and 80 mm: (a) hI for pore diameters of dpore ¼ 10, 25, and 40 nm: (a) hI for steam
steam reforming and (b) hIII for reverse methanation. reforming and (b) hIII for reverse methanation.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 11

700
C, whereas its effective binary diffusivities of CH4 with The intrinsic reaction kinetics for the SMR process sum-
H2O, H2, CO, and CO2 are 26.3  106 m2/s, 70.7  106 m2/s, marized in Eqs. (7e9) are defined per unit mass catalyst (such
21.8  106 m2/s, and 17.8  106 m2/s, respectively, at T∞ ¼ as kmol/kgcat-h), as proposed by Xu and Froment [11,12]. In the
700
C and pt;∞ ¼ 1 bar. The effectiveness factors hI and hIII are experiment, Xu and Froment [11,12] used commercial catalyst
plotted as a function of the methane conversion ratio, bCH4 , in pellets that had an apparent density of 2355 kg/m3 and con-
Fig. 10, where the mean pore diameter, dpore , is varied between tained 15.2 wt% nickel supported on magnesium spinel (Ni/
10, 25, and 40 nm. As shown in Fig. 10, larger pore diameters MgAl2O4). They also showed that the surface area of nickel in
lead to higher effectiveness factors, which can be contributed the catalyst was 9300 m2/kgcat (and the total surface area was
to the increased mass transport rate. 58,000 m2/kgcat) through BET measurements. In this study, the
Fig. 11 shows the same effectiveness factor data shown in apparent density of the washcoat layer is assumed to be
Fig. 10 with respect to the modified Thiele moduli, fI and fIII , 2355 kg/m3 (Table 1) based on the work of Xu and Froment
which are defined as follows to consider the pore diameter effect: [11,12]. Then, the effectiveness factors obtained for rcat ¼
0:9  0:6 
 0:16 2355 kg/m3 can be regarded as those obtained for the volume-
T∞ SC dpore specific catalyst surface area, Acat , of 2.2  107 m2/m3 (2355 kg/
fIII ¼ fI

;
700 C 3 25nm m3  9300 m2/kgcat ¼ 2.2  107 m2/m3 ¼ 22 mm1). Considering
 0:9  0:6  0:05 (37)
T∞ SC dpore this fact, the effect of Acat on the effectiveness factors is
fIII ¼ fIII :
700
C 3 25nm studied in Figs. 12 and 13.
Fig. 11 clearly shows that the correlation equations of Eq.
(34) can be used to determine the effectiveness factors for the
various mean pore diameters considered in this study.
0.8
0.8 Fixed @ pt , = 1 bar, SC = 3 (a)
Fixed @ pt , = 1 bar, SC = 3 (a) Catalyst Surface Area
Acat= 11 m-1: T = 600 C
I

0.6
22 m-1:
Effectiveness Factor,

Pore Diameter
I

0.6 T = 600 C 33 m-1:

dpore= 10 nm:
Effectiveness Factor,

25 nm:
40 nm: 0.4 700 C
0.4
700 C
1.5
I = , I 1 0.2
1 + 0.41( I* )1.88
0.2
800 C Correlation
800 C
0
0 0.2 0.4 0.6 0.8 1
0
0 4 8 12 16 20 Methane Conversion, CH 4
Modified Thiele Modulus, *
I 0.8
0.8 Fixed @ pt , = 1 bar, SC = 3 (b)
Fixed @ pt , = 1 bar, SC = 3 (b)
III

Catalyst Surface Area T = 600 C

0.6
Acat= 11 m-1:
III

Pore Diameter
Effectiveness Factor,

0.6
T = 600 C dpore= 10 nm: 22 m-1:
Effectiveness Factor,

25 nm: 33 m-1:
40 nm: 700 C
0.4
0.4

700 C 1.5
III = , III 1 0.2
* )1.03
1 + 4.14( III
0.2
Correlation
800 C
800 C
0
0 0 0.2 0.4 0.6 0.8 1
0 4 8 12 16 20 Methane Conversion, CH 4
Modified Thiele Modulus, *
III
Fig. 12 e Variations in the effectiveness factors with
Fig. 11 e Variations in the effectiveness factors with respect to the methane conversion ratio, bCH4 , for volume-
respect to the modified Thiele moduli, fI and fIII , for mean specific catalyst surface areas of Acat ¼ 1.1, 2.2, and
pore diameters of dpore ¼ 10, 25, and 40 nm: (a) hI for steam 3.3 £ 107 m2/m3: (a) hI for steam reforming and (b) hIII for
reforming and (b) hIII for reverse methanation. reverse methanation.

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
12 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4

0.8 Fig. 13 shows the effectiveness factors with respect to the

Fixed @ pt , = 1 bar, SC = 3 (a) modified Thiele moduli. Again, correlation equations provided
in Eq. (34) can be used to determine the effectiveness factor
data for the various Acat values considered in this study.
Catalyst Surface Area
I

0.6 T = 600 C
Acat= 11 m-1: Finally, the modified Thiele moduli for all the reforming
Effectiveness Factor,

22 m-1: conditions and washcoat property variations considered in

33 m-1: this study can be summarized as
0.4  0:18  1:3  0:07  0:37
pt;∞ T∞ SC Lcat
fI ¼ fI
700 C 1.5 1bar 700
C 3 50mm
= , 1  0:16  0:2 (39)
I
1 + 0.41( I* )1.88
I dpore Acat
 ;
0.2 25nm 22mm1
800 C Correlation
 0:2  0:9  0:6  0:15
pt;∞ T∞ SC Lcat
fIII ¼ fIII
0 1bar 700
C 3 50mm
 0:05  0:1 (40)
0 4 8 12 16 20 dpore Acat
 :
Modified Thiele Modulus, *
I
25nm 22mm1
0.8
Numerical experiments were performed to quantitatively
Fixed @ pt , = 1 bar, SC = 3 (b) determine the correlation errors for two extreme cases, i.e.,
the highest effectiveness case (T∞ ¼ 600
C, SC ¼ 2, Lcat ¼
III

0.6 Catalyst Surface Area 20 mm, dpore ¼ 40 nm, Acat ¼ 11 mm1) and the lowest effec-
T = 600 C Acat= 11 m-1: tiveness case (T∞ ¼ 800
C, SC ¼ 4, Lcat ¼ 80 mm, dpore ¼ 10 nm,
Effectiveness Factor,

22 m-1:
33 m-1: Acat ¼ 33 mm1) at pt;∞ ¼ 1, 2, and 3 bar. The correlation errors
were not significant in those extreme conditions except for
0.4 the very initial stages of the SMR process, such that relative
errors are less than 22% for hI and less than 26% for hIII for
1.5
700 C III = , III 1 bCH4 > 0:05. Thus, it is believed that one can adequately esti-
* )1.03
1 + 4.14( III
0.2 mate the effectiveness factors of washcoat catalyst layers for
Correlation the SMR process using the correlation equations of Eq. (34)
800 C
along with Eqs. (22), (39) and (40).

0
0 4 8 12 16 20
Conclusion
Modified Thiele Modulus, *
III

The effectiveness factors of the SMR process in the washcoat

Fig. 13 e Variations in the effectiveness factors with
catalyst layer were numerically investigated by considering
respect to the modified Thiele moduli, fI and fIII , for
the chemical reactions and multicomponent mass transport
volume-specific catalyst surface areas of Acat ¼ 1.1, 2.2, and
under reforming conditions relevant to small-scale hydrogen
3.3 £ 107 m2/m3: (a) hI for steam reforming and (b) hIII for
production. The intrinsic reaction kinetics proposed by Xu and
reverse methanation.
Froment [11,12] were considered. The reformate gas composi-
tion predetermined as a function of the methane conversion
ratio was applied on the surface of the washcoat layer as
boundary conditions. The effectiveness factors were investi-
Fig. 12 shows the effectiveness factors hI and hIII obtained gated for various reforming conditions and washcoat proper-
by varying the volume-specific catalyst area, Acat , between 11, ties and presented as a function of the methane conversion
22, and 33 mm1 as a function of the methane conversion ratio, ratio. In addition, simple correlation equations were developed
bCH4 . This variation was implemented by simply multiplying along with the introduction of the modified Thiele moduli. It
the intrinsic reaction rates by a scaling factor (0.5, 1.0, and 1.5) was demonstrated that the proposed correlation equations
in the actual simulations. As shown in Fig. 12, a higher could adequately estimate the effectiveness factor data ob-
volume-specific catalyst area, Acat , results in lower effective- tained by numerical calculation. Therefore, the present study
ness factors because of the corresponding increase in the re- is believed to provide useful data for designing small-scale
action rates inside the washcoat layer. The modified Thiele SMR systems and performing numerical simulations.
moduli for considering the catalyst surface area effect are
defined as
0:9  0:6 
 0:2 Acknowledgment
T∞ SC Acat
fI ¼ fI ;
700
C 3 22mm1
 0:9  0:6  0:1 (38) This research was supported by the National Research Foun-
T∞ SC Acat
fIII ¼ fIII : dation of Korea (NRF) funded by the Ministry of Education
700
C 3 22mm1
(NRF-2015R1D1A1A01057376).

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4 13

the steam reforming and water-gas shift reactions occur in

Appendix A. Summary of the reaction rate the reformer by r1 and r2 mol, respectively, then the initial,
constants from Xu and Froment [11,12]. change, and final mole number of gas species become as fol-
lows [26].
The reaction rate constants are given as
  0:5
240; 100 kmol,bar
kI ¼ 4:225  1015 exp  ; (A1)
Rg T kgcat ,h Gas Species Initial Change Final
CH4 1 er1 1er1
 
67; 130 kmol H2O SC er1er2 SCer1er2
kII ¼ 1:955  106 exp  ; (A2)
Rg T bar,kgcat ,h H2 0 3r1 þ r2 3r1 þ r2
CO 0 r1er2 r1er2
  0:5
CO2 0 r2 r2
243; 900 kmol,bar Total Moles 1 þ SC 2r1 1 þ SC þ 2r1
kIII ¼ 1:020  1015 exp  : (A3)
Rg T kgcat ,h

The adsorption coefficients are given as

The final gas composition must satisfy the equilibrium
 
4 38; 280 1 constraints provided in Eqs. (B1) and (B2), resulting in a set of
KCH4 ¼ 6:65  10 exp bar ; (A4)
Rg T nonlinear equations, written as

  f1 ðr1 ; r2 Þ ¼ 0
88; 680
KH2 O ¼ 1:77  105 exp  ; (A5)
Rg T : Keq;I ð1  r1 ÞðSC  r1  r2 Þð1 þ SC þ 2r1 Þ2

   p2t ð3r1 þ r2 Þ3 ðr1  r2 Þ

82; 900 1
KH2 ¼ 6:12  10 9
exp bar ; (A6) ¼ 0; (B3)
Rg T

  f2 ðr1 ; r2 Þ ¼ 0 : Keq;II ðr1  r2 ÞðSC  r1  r2 Þ  ð3r1 þ r2 Þðr2 Þ ¼ 0: (B4)

70; 650 1
KCO ¼ 8:23  105 exp bar ; (A7)
Rg T These nonlinear equations are solved using the Newton-
Raphson method within 10 iterations (initial values of
The equilibrium constants are given as
r01 ¼ 0:7 and r02 ¼ 0:4 are recommended).
 
26; 830 2
Once r1 and r2 are obtained by solving Eqs. (B3) and (B4), the
Keq;I ¼ exp  þ 30:114 bar ; (A8) eq
T equilibrium methane conversion ratio, bCH4 , and the mole
eq
fraction of methane, xCH4 , are evaluated as
 
4; 400
Keq;II ¼ exp  4:036 ; (A9) eq eq 1  r1
T bCH4 ¼ r1 ; xCH4 ¼ : (B5)
1 þ SC þ 2r1
 
22; 430 2
Keq;III ¼ Keq;I Keq;II ¼ exp  þ 26:078 bar : (A10)
T references

Appendix B. Equilibrium composition calculation [1] Ridler DE, Twigg MW. Steam reforming. In: Twigg MW,
[32]. editor. Catalysts handbook. second ed. 1989. p. 225e82.
Wolfe, London.
eq [2] Larminie J, Dick A. Fuel cell systems explained. second ed.
To determine the equilibrium methane conversion ratio, bCH4 ,
eq Chichester: John Wiley & Sons; 2003.
and the equilibrium mole fraction of methane, xCH4 , the
[3] O'Hayre RP, Cha SW, Colella W, Prinz FB. Fuel cell
equilibrium constants for the steam reforming (reaction I) and fundamentals. New York: John Wiley & Sons; 2006.
the water-gas shift (reaction II) should be considered, which [4] Calo E, Giannini A, Monteleone G. Small stationary reformers
are given as [11,12]. for H2 production from hydrocarbons. Int J Hydrogen Energy
2010;35:9828e35.
 
x3H2 xCO 2 n3H2 nCO 26; 830 2 [5] Schjølberg I, Hulteberg C, Yasuda I, Nelsson C. Small scale
Keq;I ¼ pt ¼ p2
t ¼ exp 30:114  bar ;
xCH4 xH2 O nCH4 nH2 O n2t T reformers for on-site hydrogen supply. Energy Procedia
(B1) 2012;29:559e66.
[6] Dincer I, Acar C. Review and evaluation of hydrogen
  production methods for better sustainability. Int J Hydrogen
xH2 xCO2 nH nCO2 4; 400 Energy 2015;40:11094e111.
Keq;II ¼ ¼ 2 ¼ exp 4:036 þ ; (B2)
xCO xH2 O nCO nH2 O T [7] Kim TW, Park JC, Lim TH, Jung H, Chun DH, Lee HT, et al. The
kinetics of steam methane reforming over a Ni/g-Al2O3
where pt is the total pressure (bar), T is the temperature, and xi
catalyst for the development of small stationary reformers.
and ni are the mole fraction and mole number of gas species i, Int J Hydrogen Energy 2015;40:4512e8.
respectively. [8] Yang JI, Kim TW, Park JC, Lim TK, Jung H, Chun DH.
Consider a reformer operating at a fixed total pressure and Development of a stand-alone steam methane reformer for
S/C ratio with 1 mol of methane and SC mol of water vapor. If

Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
14 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 4
on-site hydrogen production. Int J Hydrogen Energy
2016;41:8176e83.
[9] Van Hook JP. Methane-steam reforming. Catal Rev  Sci Eng
1980;21:1e51.
[10] Rostrup-Nielsen JR. Production of synthesis gas. Catal Today
1993;18:305e24.
[11] Xu J, Froment GF. Methane steam reforming, methanation
and water-gas shift: I. Intrinsic kinetics. AIChE J
1989;35:88e96.
[12] Xu J, Froment GF. Methane steam reforming: II.
Diffusional limitations and reactor simulation. AIChE J
1989;35:97e103.
[13] Agar DW. Multifunctional reactors: old preconceptions and
new dimensions. Chem Eng Sci 1999;54:1299e305.
[14] Karim A, Bravo J, Gorm D, Conant T, Datye A. Wall coating of
a CuO/ZnO/Al2O3 methanol steam reforming catalyst for
micro-channel reformers. Catal Today 2005;110:86e91.
[15] Dokamaingam P, Assabumrungrat S, Soottitantawat A,
Sramala I, Laosiripojana N. Modeling of SOFC with indirect
internal reforming operation: comparison of conventional
packed-bed and catalytic coated-wall internal reformer. Int J
Hydrogen Energy 2009;34:410e21.
[16] Seo YS, Seo DJ, Seo YT, Yoon WL. Investigation of the
characteristics of a compact steam reformer integrated with
a water-gas shift reactor. J Power Sources 2006;161:1208e16.
[17] Kuroki M, Ookawara S, Ogawa K. A high-fidelity CFD model
of methane steam reforming in a packed bed reactor. J Chem
Eng Jpn 2009;42:s73e8.
[18] Jung UH, Kim W, Koo KY, Yoon WL. Genuine design of
compact natural gas fuel processor for 1-kWe class
residential proton exchange membrane fuel cell systems.
Fuel Process Technol 2014;121:32e7.
[19] Zanfir M, Gavriilidis A. Catalytic combustion assisted
methane steam reforming in a catalytic plate reactor. Chem
Eng Sci 2003;58:3947e60.
[20] €
Simsek E, Avci AK, Onsan ZI. Investigation of catalyst
performance and microstructured reactor configuration for
syngas production by methane steam reforming. Catal
Today 2011;178:157e63.
[21] Mundhwa M, Thurgood CP. Numerical study of methane
steam reforming and methane combustion over the
segmented and continuously coated layers of catalysts in a
plate reactor. Fuel Process Technol 2017;158:57e72.
[22] Roh HS, Lee DK, Koo KY, Jung UH, Yoon WL. Natural gas
steam reforming for hydrogen production over metal
monolith catalyst with efficient heat-transfer. Int J Hydrogen
Energy 2010;35:1613e9.
[23] Yang JI, Ryu JH, Lee KY, Jung NJ, Park JC, Chun DH, et al.
Combined pre-reformer/reformer system utilizing monolith
Please cite this article in press as: Jeong A, et al., Effectiveness factor correlations from simulations of washcoat nickel catalyst layers for
small-scale steam methane reforming applications, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/
j.ijhydene.2018.06.059
catalysts for hydrogen production. Int J Hydrogen Energy
2011;36:8850e6.
[24] de Miguel N, Manzanedo J, Arias PL. Testing of a Ni-Al2O3
catalyst for methane steam reforming using different
reaction systems. Chem Eng Technol 2012;35:720e8.
[25] Palma V, Miccio M, Ricca A, Meloni E, Ciambelli P. Monolithic
catalysts for methane steam reforming intensification:
experimental and numerical investigations. Fuel
2014;138:80e90.
[26] Fogler HS. Elements of chemical reaction engineering. fourth
ed. New Jersey: Prentice Hall; 2006.
[27] Rawlings JB, Ekerdt JG. Chemical reactor analysis and design
fundamentals. second ed. Madison: Nob Hill Publishing;
2011.
[28] Elnashaie SSEH, Abashar MEE. Steam reforming and
methanation effectiveness factors using the dusty gas model
under industrial conditions. Chem Eng Process
1993;32:177e89.
[29] Alberton AL, Schwaab M, Fontes CE, Bittencourt RC, Pinto JC.
Hybrid modeling of methane reformers. 1. A metamodel for
the effectiveness factor of a catalyst pellet with complex
geometry. Ind Eng Chem Res 2009;48:9369e75.
[30] Pantoleontos G, Kikkinides ES. Georgiadis MC). A
heterogeneous dynamic model for the simulation and
optimisation of the steam methane reforming reactor. Int J
Hydrogen Energy 2012;37:16346e58.
[31] Baek SM, Kang JH, Lee KJ, Nam JH. A numerical study of the
effectiveness factors of nickel catalyst pellets used in steam
methane reforming for residential fuel cell applications. Int J
Hydrogen Energy 2014;39:9180e92.
[32] Nam JH. Effectiveness factor correlations for spherical nickel
catalyst pellets used in small-scale steam methane
reformers. Int J Hydrogen Energy 2015;40:5644e52.
[33] Stutz MJ, Poulikakos D. Optimum washcoat thickness of a
monolith reactor for syngas production by partial oxidation
of methane. Chem Eng Sci 2008;63:1761e70.
[34] Mason EA, Malinauskas AP. Gas transport in porous media:
the dusty gas model. New York: Elsevier; 1983.
[35] Krishna R, Wesselingh JA. The MaxwelleStefan approach to
mass transfer. Chem Eng Sci 1997;52:861e911.
[36] Todd B, Young JB. Thermodynamic and transport properties
of gases for use in solid oxide fuel cell modeling diffusion
viscosity SOFC thermal conductivity SOFC modelling gas
mixture properties. J Power Sources 2002;110:186e200.
[37] Fuller EN, Schettler PD, Giddings JC. A new method for
prediction of binary gas-phase diffusion coefficients. Ind Eng
Chem 1966;58(5):19e27.
[38] Klein SA. Engineering equation solver (EES), ver. 9.254. F-
chart software. 2012.